Microsoft is looking at next-generation nuclear reactors | Hacker News Reader

Microsoft is looking at next-generation nuclear reactors

theverge.com

116 pointsmoneil97187 comments

I just recently read this very well-written article on why SMRs don't make sense (I don't recall, but I probably got the link here!).

https://cleantechnica.com/2023/11/30/what-drives-this-madnes...

sample excerpt: To be clear, if SMRs made sense, existing nuclear power generation facilities are the place to build them. They are already at the centre of the seven overlapping layers of defence that nuclear generation sites require from the international, all supply and waste chains and the physical and electronic security of the facility itself.

It is a maddeningly poorly written article in my opinion.

Terribly long waffling about small and rather unconvincing human problems and politics.

I kept hoping that he will get to actually spelling out the x, y, and z reasons why SMR's are not viable.

Instead I got one of those ultra long sales pitches that has a wall of text but no argumentation.

Currently I'm in a position where I don't buy any of the marketing SMR people put out before they start delivering projects and start hitting economic and reliability targets.

I would have wished a conclusive and convincing argument "this is the reason why SMR makes as little sense as hydrogen" but it just was not in that article.

EDIT: he points out Bent Flyvbjergs work about megaprojects. I highly recommend his latest book on delivering and failing to deliver Megaprojects[0].

[0] https://books.google.sc/books?id=95RCzwEACAAJ&q=How+Big+Thin...

I'll give that a read too, thanks.

One of that article's points was that nuclear plants only succeeded as mega projects, because: 1) they dovetailed with weapons programs so had a huge additional funding pool, 2) big projects add thermal efficiencies that bring the cost per kwh down significantly 3) any project needs near military grade security, which doesn't doesn't suit small sites

Ontario will be building an SMR near an existing large-scale nuclear facility. The promo mock-up picture of the SMR shows a Canada Post van driving right up.

This is a little absurd, given that the road going up to similar plants have a number of warning signs to turn around if you have no business going there. This includes a very ominous “last chance” sign, probably before paramilitary forces pull you over.

I ran into this situation recently when I naïvely went looking for the Nuclear Information Centre (a sort of nuclear mini-museum) that used to be near the gates of Darlington about 25-30 years ago.

(1) https://www.opg.com/wp-content/uploads/2023/06/BWRX-300-rend...

(2) https://www.opg.com/news-resources/education/visitor-centres...

Is the thermal efficiency really relevant? I always understood the externalities to drive much of the costs.

And even if you're building a mega site with military defences, you could still build it by having 10 small reactors. If they are cheaper to build "per megawatt" due to economies of scale, you are still ahead.

Here is the article by the same author where he actually lays out the case against SMRs. https://cleantechnica.com/2021/05/03/small-modular-nuclear-r... Fairly convincing IMO.

Incorrect headline. More accurate would be: "Microsoft trying to hire someone to investigate nuclear power". This is the first in a series of about 100 steps before they actually build a reactor. And the overwhelming likelihood is that they study it, decide it's too risky/expensive, and shelve the project.

Nuclear is a reasonable source of green power, but this is the fourth-or-so time I've seen a MAANG suggest they're going to use it. I'd ignore this until someone breaks ground.

*breaks atom

MANGA?

(Or would that be the alternative history version where the companies were Japanese)

MAANG is pretty established, because it just follows from FAANG and Facebook->Meta.

I've never seen it before. And MANGA is much funnier and more memorable. Netflix seems like it should be ditched in favor of Microsoft though.

Also haven't seen it before. MAGMA would be a pretty sweet abbreviation!

It’s a pretty common acronym

Microsoft Apple Google Meta Amazon

You have to pronounce "MAANG" like Tony Montana would.

If we're gonna just arbitrarily name enshitified companies I propose the following instead:

Facebook

Ubisoft

Chrysler

Kroger

Enron

Verizon

Exxon

YouTube

TikTok

Twitter

Intel

Netflix

Google

Oh I see what you did there. Very accurate acronym from that list.

Somewhat confusing when GP uses it to refer to Microsoft, though.

It seems like a really dumb idea to use the most expensive form of electricity generation by some margin for a very predictable work load.

Most expensive if you discount reliability, longevity and persistence in energy generation.

A predictable energy generation source for a predictable work load makes sense.

Nuclear fans grossly overstate the cost of dealing with renewable intermittency, especially if the load has dispatchability. Servers don't necessarily have to run 24/7.

> Servers don't necessarily have to run 24/7.

I'm curious, do you actually run servers for a living? I'm wondering what a typical use case is (aside from crypto mining) where powering off and on physical servers due to demand, cost, or availability of power is a thing.

In fairness, training machine learning models is definitely one of areas where only running when it’s cheap to run makes some amount of sense. It’s something you want done eventually, not something that is responding to an immediate need.

In practice…yeah. The machine’s your running your models on are expensive, the people working on the model waiting extra time to check the new model are expensive, and not keeping up with your competitors because you only work half the time is not generally a wise choice for someone like Microsoft.

I have worked for a smaller company that trained a model on AWS when spot instances were cheap…until they got large enough to purchase some hardware best suited to their use case that could run almost constantly, then doing everything except training their model in the cloud. So yeah I also doubt this makes sense in practice.

The cost of supplying the servers with renewables 90% of the time would be much cheaper than supplying them 100% of the time. The former can easily be done with some batteries; the latter requires more extensive consideration of rare dark-calm periods.

Another consideration is that servers can be located in the best places for renewables, not in places with high seasonal variability and relatively low winds.

> Servers don't necessarily have to run 24/7.

The benefit of your services having 25/7 availability aside, if the capital costs are high compared to operational costs, reducing operational time to save on operation costs is a bad trade.

"The wind is blowing! Quick, everyone run your Hadoop jobs!"

-- A++ customer experience

If you invest a lot of money into very expensive Nvidia training hardware you certainly want to run them as close to 24/7 as possible.

Dispatchable load usually means oversizing the dispatchable consumer to get the same overall output. This is already uneconomical for even particularly energy intensive industries (e.g. aluminum smelting). I would assume server hardware is a lot more capital intensive than that.

While I can’t speak to aluminium specifically, smelters in general have the problem that they’re pumping molten metal around. And the thing that keeps it molten inside the pipes is, usually, just inertia.

Pausing the plant doesn’t just mean pushing a button. Safely shutting it down is a long process of draining every single part of molten metal first, and might not even be doable; almost any other factory would have an easier time.

Molten cryolite rather; I think the metal pooled in the bottom of a pot could be drained off or even allowed to solidify (it would still be conductive.)

A related problem is that the walls of the pot cannot be allowed to get too hot. If this were not the case, they could simply insulate the pots to reduce heat losses to some arbitrarily low level.

What you think that will accomplish? We already schedule industrial loads for times of low demand from residential and commercial. Rescheduling those loads to coincide with high intermittent generation usually requires more transmission and distribution capacity.

Debating total cost depends almost entirely on the particular grid in question. For all we know, those "nuclear fans'" estimates could be spot-on but completely irrelevant to you.

No, but surely Microsoft could work out a deal with local utility companies to sell any excess electricity they generate?

Fossil fuel plants are far and away the most expensive form of electricity generation, we just allow them export a big chunk of the costs onto other people so we can pretend it's cheap.

Citation needed

Okay. Per [1] the cost of 1 ton of CO2 emitted is about $3,000. Per [2] fossil fuel plants emit about 1-2.3 pounds of CO2 per kWh, or about 0.0005 to 0.0012 tons per kWh. Multiplying that out (0.0005 * 3000) yields about $1.50 to $3.60 per kWh, accounting only for the cost of emissions.

Per [3] nuclear is about $0.07 per kWh. That's about 20-50 times cheaper than fossil fuel plants.

It would be fair to round that up a bit for construction and fuel extraction and refinement emissions, but don't forget to do that for fossil fuel plants, too.

[1] https://en.wikipedia.org/wiki/Social_cost_of_greenhouse_gas_...

[2] https://www.eia.gov/tools/faqs/faq.php?id=74&t=11

[3] https://css.umich.edu/publications/factsheets/energy/nuclear...

Source [1] shows $300 not $3000, which changes the math quite a bit.

China and India are building nuclear power plants pretty cheaply.

China's nuclear plans are consistently not being met. India's nuclear output is quite small, and is already exceeded by a factor of 2 by their PV output.

My point was that both are finishing several plants per year, with reasonable cost ($3k per kW compared to $8-12k per kW in the US) and within reasonable time frames of 5-10 years of construction. This is competitive with renewables if you take the cost of storage into account.

They are not competitive with renewables, which is why China and India are building so much more renewable capacity than they are nuclear capacity.

Ultimately, storage will be needed. But at that point, nuclear becomes even less desirable in those countries, because renewables are creating long periods where there's no demand to soak up the nuclear plants' output. You might object that this is a renewable problem, but I'd retort it's a nuclear problem -- they can't save much money by curtailing output, and this inflexibility is a blemish on nuclear.

With reasonable projections of the cost of storage, PV will end up much cheaper than nuclear in India, btw. India's a great place for PV; they could go full renewable with just PV and batteries, not needing wind or hydrogen.

Even with nuclear, ultimately storage will be needed. Throttling production of such an expensive piece of capital with low operational costs doesn't make much sense, it better to build to the average power needed and smooth out the peaks and valleys with a battery.

But if you have a battery, it makes more sense to fill it with low LCOE renewables rather than nuclear. So adding a battery cannot provide much revenue to the nuclear plant.

I agree, I just think it is funny when you have a large majority of your power coming from renewables or nuclear, you face similar looking problems. You are faced with the choice of artificially lowering a nuclear plants capacity factor by throttling down, or using batteries to keep the capacity factor high.

The price of grid batteries in the future will basically fully determine what power generation technologies we use.

China completes multiple reactors per year, but India does not. Here are the power reactor statistics for both countries from the International Atomic Energy Agency's Power Reactor Information System (sort by date of first grid connection):

India has connected 3 reactors to the grid since 2013-01-01:

https://pris.iaea.org/PRIS/CountryStatistics/CountryDetails....

China has connected 39 in the same time period:

https://pris.iaea.org/PRIS/CountryStatistics/CountryDetails....

China is barely building nuclear anymore. China added more wind and solar the past nine months than all of its nuclear reactors under construction will provide. Yes, that includes capacity factor.

https://twitter.com/yo_ean/status/1718633487454904718

Great to see Microsoft deploying some of it's capital to develop SMRs, the potential seems great, but governments don't seem to have gone further than nuclear submarines.

Zero capital deployed

Title is vague and editorialised.

"Microsoft plans to power AI data centers with next-generation nuclear reactors" would be better.

I built HackYourNews.com (free) to tackle exactly this. Here's how this story showed up there:

Dehyped title: Microsoft considers nuclear energy including small modular reactors to power data centers and AI ambitions due to energy demands.

Summary:

Microsoft is exploring using nuclear energy to power its data centers and AI ambitions. They posted a job listing for a principal program manager to lead their nuclear energy strategy. They are specifically interested in small modular reactors, as they are supposed to be easier and cheaper to build than large traditional nuclear plants. However, small modular reactors would still face challenges like developing domestic supply chains for nuclear fuel and long-term storage of nuclear waste. Microsoft did not provide details on their nuclear plans or how they would address these issues. While nuclear energy does not produce greenhouse gases, there are debates around its role in addressing climate change and managing radioactive materials.

Comments:

Microsoft is considering using nuclear power to provide a reliable source of energy for its AI workloads. Some commenters expressed skepticism that this plan will actually be implemented, noting that similar announcements have been made in the past without follow through. Others argued that nuclear could provide a predictable energy source to match Microsoft's predictable energy needs. There was also a discussion around the relative costs of different power sources, with some claiming fossil fuels impose hidden long-term costs while nuclear has reliable long-term energy generation.

I have an idea. Why don't we give new experimental AI systems their own off-grid nuclear power plants so they can't be switched off. There's no way that could go wrong.

I can’t let you interrupt my fine-tuning run, Dave.

Colossus irl

Easy, just broaden the responsibilities of the SCRAM officer. Within reach of their axe you have:

]The cable for the control rods

]The power feed from the plant to the AI datacenter

]The backbone fiber connecting the DC to the wider world

In event of Skynet episode, sever all three!

Breaking: Guy at big company had an idea

>Microsoft has also made an audacious deal to purchase electricity from a company called Helion that’s developing an even more futuristic fusion power plant.

If we're going to 'solve' global warming, we really need fusion ASAP. It truly is a silver bullet for this problem and will also be the key to colonizing Mars.

> If we're going to 'solve' global warming, we really need fusion ASAP. It truly is a silver bullet for this problem and will also be the key to colonizing Mars.

Not at all. People have this impression that fusion produces much more energy than fission. That's not the case.

Fission produces about 200 MeV per single fission event. Since the nucleus that splits is either U-235 or Pu-239, that is about 1 MeV per nucleon.

Fusion produces in the range of 1 MeV to 23 MeV per even. There are numerous possible reactions, from Deuterium-Tritium to Hydrogen-Nitrogen. At most you get about 4 MeV per nucleon (for the Deuterium-Tritium and Deuterium-Helium3). In most cases you get less, even in some cases less than 1MeV per nucleon. So, at most fusion is about 4 times more efficient than fission. But in those cases when this happens, one of the reactants is something that we can't find in nature (either tritium or He3), so we need to make out of other nuclear reactions. Tritium in particular is something regularly discarded by fission reactors. The problem is that if we ever develop Deuterium-Tritium fusion, our fission reactors will not generate anything close to the needed Tritium, not by 2 orders of magnitude.

This is just to dispel the illusion that if we unlock fusion, we get loads more energy than from fission.

Now, it will be cleaner. It will not produce nuclear waste. But nuclear waste is not really the big issue that the anti-nuclear crowd makes it to be.

The problem with fusion is this: we need two miracles to happen. We need to make fusion happen, and then we need to make it economical.

With fission, we only need the second miracle.

The idea is if you can make one fusion generator, you can make one million, as they're safe enough to build into every city and town in the world, just like gas and coal power plants are now. The only real danger to the population will be the high voltage power lines.

You have a point.

But here's a trivia question: how many coal power plants and how many nuclear power plants are in the US?

Answer: 219 coal and 54 nuclear.

Dupe from recent past.

https://news.ycombinator.com/item?id=37627697

What happened to those thorium small scale reactors that Bill Gates was funding research on?

Their last news https://www.terrapower.com/terrapower-and-enec-announce-mou-...

Does it worth it , though? They'll cover their expenses in how many years?

Perhaps the AI can then figure out what to do with the nuclear wast....

Nuclear waste cam be sealed in casks and dropped near the ocean subduction zone. In a few thousand years it will be sucked down into the mantle.

The Russians did this before most of us were born.

It’s gonna be a long cold AI Nuclear Winter

Perhaps we just shouldn't do this?

The environmental impact of AI is already huge and we have an uncertain energy future.

Why not? By that logic, humans are a much larger impact and we should depopulate them.

This reduces the environmental impact of AI, and they can use the power plants to power something else in the future.

Or perhaps we should. Right now, an unprecedented amount of money is being raised for AI.

If we can somehow turn that into green energy, that’s a big win for the planet. Those nuclear plants will be around a lot longer than the AI “gold rush”.

No. We should do more like this if “we have an uncertain energy future”.

I don't understand?

> The environmental impact of AI is already huge and we have an uncertain energy future.

This seems like an argument FOR them doing what they're proposing.

Innovation is the only way we'll get out of this. Sitting on our asses consuming food for no innovation and a culture of shutting down any attempt like that will only make things worse. This is a good tradeoff for the environmental cost and risks which nuclear brings (which isn't even large).

On the contrary, nuclear power is the greenest energy generation source (in terms of CO2/kWh over the full lifecycle), by far.

https://www.nrel.gov/docs/fy21osti/80580.pdf

Unless you consider the CO2 emitted while you wait for the nuclear plant to come online and displace fossil fuels.

Lifecycle emissions also assume all the parts are being made with current energy mix. In the zero fossil fuel future, where is the CO2 emission for PV coming from? Concrete production? PV uses much less concrete than nuclear per levelized unit of power.

Panels don’t spring from the ether. From mining the silica and other materials that go into PV, to transporting it, refining it, transporting it again, manufacturing the panel and transporting them again across the world (since this all happens in Asia). I too am surprised by the claim that nuclear outperforms its way better and I’m a strong believer that nuclear is the only thing that can get us closer to net 0 than any other.

Here’s a link comparing lifecycle co2 and it does show that nuclear is better: https://www.carbonbrief.org/solar-wind-nuclear-amazingly-low...

The answer, of course, is to do both. PV is great and we should be building massive amounts of it. Nuclear is also great and we should also be building massive amounts of that. They each have pros & cons.

No "of course" about it. Nuclear is struggling to find any reasonable use case. Nuclear and renewables not complementary; you can't economically back up renewables with nuclear.

I'm not aware of any existing or even near-future feasible renewable solution for indoor heating over the winter months in northern climates, other than nuclear.

Nuclear exists, it's proven, it works. And it's not even that expensive. We could build two of the famously overpriced Georgia nuclear plant for the price one guy paid to shut down a social media website. We could build three of them for the price Microsoft paid to buy a video game company. And we could definitely get prices down if we actually wanted to. I think it's really, really bad policy to be quibbling over pennies when discussing solutions to this literally civilization-ending catastrophe that is staring us in the face. We have to do both. If you're worried about economics, it's way more expensive not to be building out every solution we have available to us.

Nuclear is a good solution in many cases. It does us no good to just take it off the table, at least until we actually have better solutions.

> I'm not aware of any existing or even near-future feasible renewable solution for indoor heating over the winter months in northern climates, other than nuclear.

Make hydrogen during the summer (or wind peaks). Store it underground (Europe has space in salt formations for many PWh of hydrogen, for example), use in winter to generate power during dark-calm periods, drive heat pumps with the power.

This is likely to be cheaper than a system that uses nuclear power plants to drive those heat pumps.

Thanks for your reply. There's a good response to that plan here[1], the main takeaways being low round-trip efficiency (lose 50-80% of the energy spent in the process of storing & releasing) and the lack of proven renewable hydrogen generation tech. We should absolutely keep developing this idea and use it in places where it's a good fit. But I'm not willing to bet our future on unproven tech when we have a perfectly good solution right in front of us already working today.

[1] https://theness.com/neurologicablog/hydrogen-takes-another-h...

Neither of those is a good argument. The low round trip efficiency doesn't matter much, because hydrogen is mostly there to deal with unusual events (dunkelflauten). Most of the energy still goes either directly from PV/wind to the grid, or through efficient short term storage. Viewed another way: in a storage system, the "cost of inefficiency" is proportional to the number of charge/discharge cycles; for seasonal or rare event backup the number of such cycles is much smaller than for daily storage.

Lack of proven renewable hydrogen is an argument that would also rule out nuclear. Nuclear fans are constantly pointing to new, cheaper, more wonderful systems, none of which are proven. Even building reactors more cheaply in the west involves a large leap of faith. What we do know is that electrolyzer prices have been falling rapidly, just like renewable generation and battery storage prices have been falling. They're another technology that has large numbers of similar units and has a strong manufacturing experience effect.

> Lack of proven renewable hydrogen is an argument that would also rule out nuclear.

I don't think so. We just built a new plant in Georgia. Minnesota's power grid is already 1/4 nuclear. We've done it before, we can do it again. It would be nice to get costs down, but it's not a requirement. The only thing stopping us is political will and US's general inability to accomplish big things. Hydrogen storage isn't here today (but I hope it will be soon!), nuclear is.

Reply on news.ycombinator.com